The broad objective of this proposal is to establish a computational platform to predict and understand protein functions at multi-scales, from molecule to cell, through integrating data from structural genomics, functional genomics and chemical genomics, and using tools derived from bioinformatics, chemical informatics and biophysics. There is a serious need to meet this objective in an era of proteomics where proteins are easily isolated, yet not so easily functionally classified with any degree of confidence. As the first step we propose here to: (1) design and implement scalable, accurate, reliable and robust algorithms and associated software for predicting, comparing, searching, and classifying protein functional sites and proteinligand interactions;(2) design and implement an ontology-driven protein functional site and protein-ligand interaction database that integrates comprehensive structure, function, and mutation information and supports quantitative modeling of protein structure and functions at the genome scale;(3) Establish first an intuitive graphical user interfaces (GUI) for scientists to visualize, analyze and mine functional site information for comparative proteomics and second establish an application programming interface (API) for programmers to develop new algorithms and applications using the foundation proposed here. The results of this effort will be disseminated through the Protein Data Bank which is used by over 10,000 scientists every day.